1 to 5 power divider working in 3 GHz

[yykk]

Hi all,


I wnat to design a one to five power divider working in 3 GHz, it must have a good isolation on each port. ( mircostrip structure is more prefer)

thanks


yykk

 

[tapa]

You can use some of the simetrical structures similar to "rat-race". But ustrip is the tecnology with less isolation. Yo can have a look at any uwave book.

 

[Needbias]

how to improve isolation?

 

[tapa]

Isolation?? Realy the isolation depends on circuit precision and technology. The isolation in increasing technology is:
- ustrip
- stripline
- coax (guide)

But this is for nearly all circuits, couplers, ....

 

[AL0117]

you can choose from different types of power dividers / couplers. Some of them are narrow bandwidth, some of them have better isolation. In fact is a compromise between a lot of parameters. It depends on your application
You can find some good articles on the Ansoft site :
hxxp://www.ansoft.com/about/academics/training_materials.cfm
(x=t)

I made some wilkinson type power dividers(1/2) for the 2.2 - 2.7 GHz band, I got insertion loss of 3.5-3.6 dB and isolation27dB@2.2GHz, 43dB@2.4GHz and 21dB@2.7GHz. For my application is was enough.

Cheers, Al

 

[yykk]

Thanks all of you to give me a valuable suggestion... anyone have a example for 5 ports, because even ports is easy to design, but odd number ports ...........

 

[AL0117]

The Wilkinson type power dividers can be generalized for n way power dividers, but the balancing resistors needs crossover, and in microstrip the fabrication is difficult. Maybe in LTCC technology is easier.
In which technology your divider will be fabricated?

regards, Al

 

[ajar2dtn]

Have a look at this project incorporating wideband Wilkinson.
**broken link removed**

Also has ADS files

 

[AL0117]

If you are still interested , here is an article about wilkinson type power splitters:

**broken link removed**

Also try :
**broken link removed**

there are some links too.


regards, Al

 

[flatulent]

several options

You have several options. One is to have a tree structure of wilkinsons. First split 20%-80%. Then take the 80% branch and split in two and take each of these and split them in two.

Another way is to have a tandem sequence of directional couplers tapping off a main line. This is more difficult to get right because of the coupling factor being changed by etching errors on the copper removal.

 

[No one]

Hello;
"ajar2dtn" introduced a link. I tried, but I couldn't enter the link.
No one

 

[ajar2dtn]

MINT project has moved
Try
**broken link removed**

or search for MINTWeb on google